1 paper
Chiraag Kaushik, Ran Liu, Chi-Heng Lin +5
Classification models are expected to perform equally well for different classes, yet in practice, there are often large gaps in their performance. This issue of class bias is wide…